In order to realize the intelligent online monitoring of grate faults in the sintering machine of steel plant

a grate fault image recognition method based on OpenCV image processing technology is proposed

that is

the number of grate bars

gap width

slope and the proportion of gap sticking are detected by adopting the grate structural feature recognition algorithm to characterize the fault types such as the inclination

fracture and gap sticking of grate bars. In this paper

the grate fault of No. 1 sintering plant of Xiangtan Iron and Steel Co.

Ltd. is taken as the online monitoring object. Firstly

the grate images are spliced and binarized adaptively. Then

the grate contour is extracted in the region of interest (ROI) area of the grate image and the number and gap width are counted; Then Hough transform method is used to fit the line of the image

and the line segment where the grate is located is selected. The slope of the grate is calculated by using the two endpoints to obtain the gradient of the grate. Then

the original binary image and the image processed by using the open operation are bitwise and calculated to get the location of the stuck object and its proportion. Finally

to use the multiple opening operations extract the broken position of the grate bar. The experimental results show that the error of the grate gap width detected by using the algorithm is within 4 mm and the error of the inclination rate is within ± 1°

and the grate fault can be detected quickly and effectively.